What Is llms-full.txt?
llms-full.txt is an optional, single Markdown file that contains the full text of a website’s content in one place, instead of just linking to it. Where a regular llms.txt file acts as a short guide — a summary plus links to detailed pages — llms-full.txt puts the detail itself into the file, so an AI assistant can read everything in one load.
Think of it this way: llms.txt is the table of contents, and llms-full.txt is the whole book. For a small blog, the full version is usually unnecessary weight. For a documentation site or knowledge base, it can be the file that lets an AI answer detailed questions without fetching dozens of pages one by one.
How Is llms-full.txt Different From llms.txt?
The difference is simple: llms.txt points, llms-full.txt contains. The regular file stays small on purpose, so it fits easily into an AI’s context, and the detail lives behind its links, fetched only when needed. The full file accepts a much larger file in exchange for having everything immediately available.
| llms.txt | llms-full.txt | |
|---|---|---|
| What it holds | Title, summary, and links to detailed pages | The full text of those pages, inline |
| Typical size | Small — a few kilobytes for most sites | Large — grows with every page included |
| How an AI uses it | Reads the guide, then follows needed links | Loads the whole file once and answers from it |
| Best for | Blogs, marketing sites, small business sites | Documentation, API references, help centers |
| Main risk | Several follow-up fetches may be needed | Size, staleness, and cost |
The two files are companions, not rivals. The short file stays the entry point, and the full file is linked from it for agents that want the complete text. New to the short file? Start with our guide, What Is an llms.txt File? Complete Guide for Website Owners (2026), and our llms.txt example template. Platform guides cover WordPress, Shopify, and Squarespace.
Is llms-full.txt Part of the Official Specification?
No — and this is the honest answer most explainers skip. The specification at llmstxt.org, currently at version 2, defines the llms.txt file: an H1 title, an optional summary blockquote, and H2 sections containing lists of links. The v2 text does not define a separate llms-full.txt standard.
llms-full.txt is a community convention that grew up around that proposal. Some documentation tools and generators produce a full-text variant on their own, and different tools have used slightly different names for it. That is why it is described as an optional companion file rather than a required part of the spec.
This does not make it useless, but it changes how you should treat it. There is no standards body that will punish you for skipping it, no ranking boost for having one, and no guarantee every AI tool will look for it by that exact name. Treat it as a practical option for the right kind of site, not a compliance requirement. Note too that llms.txt itself is a proposal for helping AI systems understand a site, not a ranking factor — see llms.txt vs robots.txt for how it differs from robots.txt’s access-control job.
When Does the Full Version Actually Help?
The full version earns its size when an AI needs to answer questions from your content in a single pass, without a chain of follow-up fetches. That pattern shows up in a few clear cases.
Documentation and API references. This is the strongest case. A coding agent fixing an API call benefits from having the whole reference — endpoints, parameters, examples — in one place, rather than discovering the right page through a link list first. The v2 proposal notes that software documentation is where the format is used most heavily.
Knowledge bases and help centers. If your site exists to answer questions — troubleshooting guides, policy explanations, product manuals — a full file lets an agent search the complete text at once. Fetching one help article at a time works, but it is slower and more likely to miss the article containing the answer.
Single-load agent workflows. Some tools take one context file and work from it. If your content is consumed that way — pasted into a chat, loaded into a coding assistant, bundled with a project — a full file is far more useful than a link list, because links to live pages may not be followed in that workflow at all.
When Is a Short llms.txt Enough?
For most websites, a short llms.txt is enough. If any of these describe your site, stop at the short file.
- Your site is small. A blog with twenty posts, a portfolio, or a local business site does not need a full-text dump; the short file already points an agent at everything worth reading.
- Your content changes often. News, pricing, and frequently updated guides go stale fast. An unregenerated full file confidently serves outdated text, while a short file of links ages far more gracefully.
- Your pages are the product. If your site’s value is interactive tools, calculators, or videos, a text dump captures little of it. Point agents at the right pages and describe what each one does.
- Nobody asks detailed questions about your content yet. A full file solves a retrieval problem. Build the short file first, and revisit the full version when that need appears.
The Size Trade-Off You Cannot Ignore
The size trade-off has three parts: context, cost, and freshness. AI models read a limited amount of text at once, called a context window, and the v2 proposal warns that context windows are still too small for most websites in their entirety — which is why the short file keeps detail behind links. A full file for a large docs site can be enormous, defeating its own purpose if it is too big to load conveniently.
Cost follows the same logic: every token an agent loads costs time and, in paid API workflows, money. A full file loaded in full for every question, even simple ones, wastes both. Freshness is the third problem: a full file is a snapshot, and the moment you edit a page it is out of date until regenerated. If regeneration is not automatic, it will drift quietly, with no error message.
The practical answer to all three is scoping: one full file per major section — for example, per documentation area — generated from the same source as your pages, on the same build step, so the two cannot drift apart.
How to Build an llms-full.txt File
Building a full file is straightforward if you already have a short llms.txt, because the short file already tells you what to include and in what order.
1. Start from your llms.txt. Use its link lists as your table of contents, and keep that order in the full file.
2. Gather the Markdown version of each linked page. The full file should contain clean Markdown, not raw HTML. Many documentation platforms generate Markdown page versions automatically, and the v2 proposal recommends offering one at the same URL with a .md extension.
3. Keep the same H1 and summary. Open with the site name as the H1 and the same summary blockquote, so the two files are recognizably the same document in short and long form.
4. Put each page’s content under its own H2 section. Use the section names from your llms.txt, keep inner headings intact, and strip navigation, footers, and ads — the file should contain content, not chrome.
5. Link the two files together. Add the full file to your llms.txt — the Optional section is a natural home — so agents starting at the short file can discover the long one.
6. Generate it, don’t hand-maintain it. Put the build step in the same process that publishes your site, regenerate on every deploy, and check the output size each time: an unexpected jump means something changed upstream.
# Your Site Name
> One-sentence summary of what this site covers.
## Getting Started
Full text of the getting-started guide, in Markdown.
## API Reference
Full text of the API reference, in Markdown.Mistakes to Avoid With llms-full.txt
- One giant file for a huge site. Scope it by section, or do not build it. A file nobody can load conveniently helps nobody.
- Including content agents should not see. Drafts, internal notes, and paywalled or private pages do not belong in a public text file just because a generator swept them in.
- Letting it go stale. A full file that contradicts your live pages is worse than no file. Tie regeneration to deployment.
- Expecting a ranking or citation boost. No evidence supports that claim. Judge the file on whether it helps agents answer questions accurately.
- Replacing llms.txt with it. Keep the short file: it is the standard entry point, defined by the proposal, that other tools already know to look for.
Frequently Asked Questions
Do I need both llms.txt and llms-full.txt?
No. Start with llms.txt, and add the full version only for documentation, a knowledge base, or a single-load workflow where complete text in one file genuinely helps.
Is llms-full.txt an official standard?
No. The llmstxt.org proposal (v2) defines llms.txt; llms-full.txt is a community convention, not a separately standardized format.
Where should llms-full.txt live?
At your site root, next to llms.txt — for example, https://example.com/llms-full.txt — or under the path it covers, such as /docs/llms-full.txt. Link to it from your llms.txt.
How big is too big?
There is no official limit. If loading the whole file is slow, expensive, or likely to exceed the context an agent will devote to one question, split it by section.
Can I generate it automatically?
Yes, if your docs platform or generator supports a full-text export — run that step on every deploy so the file never drifts from your live pages.
Does llms-full.txt help SEO?
There is no evidence it improves search rankings. Its value is practical: helping AI assistants answer detailed questions from your content accurately.
Conclusion: Start Short, Go Full Only With a Reason
llms-full.txt is a tool for a specific job: giving an AI the complete text of a documentation set, knowledge base, or reference in a single load. For everyone else — blogs, marketing sites, small business sites — the short llms.txt already does the job, with a fraction of the size, cost, and maintenance burden.
So publish a good llms.txt first, keep it current, and add a full version only when you can name the questions it will answer better than the short file does. For the short file itself, try our free llms.txt generator, and compare tools in 7 Best Free llms.txt Generators, Tested and Compared (2026).
Last updated: October 2026